There are a number of optional and required dependencies necessary when building different components of OpenVDB. Below you'll find a full list of all dependencies, any minimum/recommended version requirements and some guides for getting these dependencies installed. You'll also find information on any known issues with particular software. Please report any issues not listed here to the OpenVDB TSC.

Note that OpenVDB is production tested against the VFX Reference Platform. Not all dependencies exist in the VFX Reference Platform, but for those that do, their specified versions are always recommended.

OpenVDB Components

Component	Requirements	Optional
OpenVDB Core Library	CMake, C++14 compiler, TBB::tbb, Boost::system, Boost::iostream	Blosc, ZLib, Log4cplus, IlmBase::Half
OpenVDB Print	Core Library dependencies	-
OpenVDB LOD	Core Library dependencies	-
OpenVDB Render	Core Library dependencies	OpenEXR, IlmBase, libpng
OpenVDB View	Core Library dependencies, OpenGL, GLFW3, GLEW*	-
OpenVDB Python	Core Library dependencies, Python, Boost::python	Boost::numpy, NumPy
OpenVDB AX	Core Library dependencies, LLVM	Bison, Flex
NanoVDB	-	Core Library, CUDA, TBB, Blosc, ZLib
OpenVDB UnitTests	Core Library dependencies, GoogleTest	CppUnit**
OpenVDB Documentation	Doxygen	-

* GLEW is only required for building the vdb_view binary on Windows.
** CppUnit is only required for building the AX unit tests.
Note - Bison and Flex are only required for re-generation of the AX grammar which is not typically necessary for the majority of users.

Dependency Table

Package	Minimum	Recommended	Description	apt-get	Homebrew	Source
CMake	3.12	Latest	Cross-platform family of tools designed to help build software	Y	Y	https://cmake.org
GCC	6.3.1	6.3.1	C++ 14 Compiler: The GNU Compiler Collection	Y	Y	https://www.gnu.org/software/gcc
Clang	3.8	Latest	C++ 14 Compiler: A C language family frontend for LLVM	Y	Y	https://clang.llvm.org
Intel ICC	17	Latest	C++ 14 Compiler: Intels C++ Compiler	Y	Y	https://software.intel.com/en-us/c-compilers
MSVC	19.10	Latest	C++ 14 Compiler: Microsoft Visual C++ Compiler	Y	Y	https://visualstudio.microsoft.com/vs
IlmBase	2.2**	2.3	Used half precision floating points and EXR serialization support	Y	Y	http://www.openexr.com
OpenEXR	2.2**	2.3	EXR serialization support	Y	Y	http://www.openexr.com
TBB	2018	2018	Threading Building Blocks - template library for task parallelism	Y	Y	https://www.threadingbuildingblocks.org
ZLIB	1.2.7	Latest	Compression library for disk serialization compression	Y	Y	https://www.zlib.net
Boost	1.66	1.66	Components: system, iostreams, python, numpy	Y	Y	https://www.boost.org
LLVM	6.0.0	8.0.0	Target-independent code generation	Y	Y	https://llvm.org/
Bison	3.0.0	3.0.5	General-purpose parser generator	Y	Y	https://www.gnu.org/software/gcc
Flex	2.6	2.6.4	Fast lexical analyzer generator	Y	Y	https://github.com/westes/flex
Python	2.7	2.7	The python interpreter and libraries for C++/python bindings	Y	Y	https://www.python.org
NumPy	1.14.0	1.14.0	Scientific computing with Python	Y	Y	http://www.numpy.org
GoogleTest	1.10	Latest	A unit testing framework module for C++	Y	Y	https://github.com/google/googletest
CppUnit	1.10	Latest	A unit testing framework module for C++	N	Y	https://freedesktop.org/wiki/Software/cppunit
Blosc	1.17.0*	1.17.0	Recommended dependency for improved disk compression	Y	Y	https://github.com/Blosc/c-blosc/releases
Log4cplus	1.1.2	Latest	An optional dependency for improved OpenVDB Logging	Y	Y	https://github.com/log4cplus/log4cplus
libpng	-	Latest	Library for manipulating PNG images	Y	Y	http://www.libpng.org/pub/png/libpng.html
GLFW	3.1	Latest	Simple API for OpenGL development	Y	Y	https://www.glfw.org
OpenGL	3.2	Latest	Environment for developing portable graphics applications	Y	Y	https://www.opengl.org
GLEW	1.0.0	Latest	A cross-platform OpenGL extension loading library.	Y	Y	http://glew.sourceforge.net
CUDA	-	Latest	Parallel computing platform for graphical processing units.	Y	N	https://developer.nvidia.com/cuda-downloads
Doxygen	1.8.8	1.8.11	Documentation generation from C++	Y	Y	http://www.doxygen.nl

* See the note in known issues regarding supported blosc versions.
** IlmBase/OpenEXR 2.2 is supported as Houdini 18.0 is shipped with VFX Reference Platform 2019, except for using older versions of these libraries.

Known Issues

GLFW: Crash with GLFW 3.2.1 when used with EGL on Linux. OpenVDB View may segfault on closure. Fixed in GLFW 3.3
Boost 1.66/1.67 + Python3 + NumPy: Known issue on UNIX systems which can cause the below python runtime error. Consider updating to Boost 1.68 if you require both Python3 and NumPy support:

1 RuntimeError: FATAL: module compiled as little endian, but detected different endianness at runtime.

Blosc: OpenVDB has historically used an old version of blosc (1.5.0) to serialize .vdb files. OpenVDB is moving to a new blosc version minimum of 1.17.0. We have tested a range of blosc versions and found that the following versions are NOT compatible with OpenVDB caches written using blosc 1.5.0. Additionally these versions of blosc may generally not be compatible with OpenVDB itself. Avoid using these versions and upgrade to the new minimum of 1.17.0 where possible.
All versions from 1.11.0 through to, but excluding, 1.14.0.
Versions 1.16.0 and 1.16.1.

For more information see:

Installing Dependencies

This section details a number of ways to get you up and running with the required OpenVDB dependencies.

Using UNIX apt-get

Note:CppUnit is unavailable using apt-get. See the other package manager methods or the manual installation options.

 #!/bin/bash
 # Core library
 apt-get install cmake                   # CMake
 apt-get install libtbb-dev              # TBB
 apt-get install zlibc                   # zlib
 apt-get install libboost-system-dev     # Boost::system
 apt-get install libboost-iostreams-dev  # Boost::iostream
 apt-get install libblosc-dev            # Blosc
 # AX
 apt-get install llvm-8-dev              # LLVM
 # Python
 apt-get install libboost-python-dev     # Boost::python
 apt-get install libboost-numpy-dev      # Boost::numpy
 apt-get install python-dev              # Python
 apt-get install python-numpy            # NumPy
 # Optional
 apt-get install libpng-dev              # libpng
 apt-get install libilmbase-dev          # IlmBase
 apt-get install libopenexr-dev          # OpenEXR
 apt-get install liblog4cplus-dev        # Log4cplus
 apt-get install googletest              # GoogleTest
 # vdb_view
 apt-get install libglfw3-dev            # GLFW
 # Documentation
 apt-get install doxygen                 # doxygen
 # CUDA for NanoVDB
 apt-get install nvidia-cuda-toolkit     # CUDA

Using Homebrew

Homebrew is a package manager for UNIX systems and is an alternative to apt-get or manual installation from source.

 #!/bin/bash
 # Core library
 brew install cmake                     # CMake
 brew install tbb                       # TBB
 brew install zlib                      # zlib
 brew install boost                     # Boost
 brew install c-blosc                   # blosc
 # AX
 brew install llvm                      # LLVM
 # Python
 brew install boost-python              # Boost-python
 brew install python                    # Python
 brew install numpy                     # NumPy
 # Unit Tests
 brew install gtest                     # GoogleTest
 brew install cppunit                   # CppUnit
 # vdb_view
 brew install glfw                      # GLFW
 # vdb_render
 brew install ilmbase                   # IlmBase
 brew install openexr                   # OpenEXR
 brew install libpng                    # libpng
 # Documentation
 brew install doxygen                   # Doxygen

UNIX Manual Installations

Some dependencies may be unavailable using the above package manager methods. Below are manual installation methods for these packages. Note that you will have to provide your chosen installation locations to the OpenVDB CMake build system when building OpenVDB.

CppUnit (for AX Unit Tests)

 #!/bin/bash
 wget http://dev-www.libreoffice.org/src/cppunit-1.13.2.tar.gz
 tar -xvzf cppunit-1.13.2.tar.gz
 cd cppunit-1.13.2
 # Install to users home directory - you may customise this location
 mkdir -p $HOME/cppunit
 ./configure --prefix=$HOME/cppunit
 make
 make install